Parallel search after


How can search after can be parallel if results from the previous page help the retrieval of the next page?

but I can see in the documentation that it is parallel


The parallel refers to many ongoing requests in parallel in this paragraph, not referring to the same user/session sending one request after the other. If you have thousands of users who each are paging through results at different times and different speeds, then having to open thousands of scroll searches might require a lot of open file handles while writing on that index happens.

Hope this clarifies things a little!

1 Like

Thanks :slightly_smiling_face:!

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.